Modelado de un algoritmo para la identificación de problemas jurídicos y tesis en las providencias de la sección primera del consejo de estado usando técnicas de procesamiento de lenguaje natural.

Abstract

The Council of State is the closing body and the highest Court of Administrative Litigation in Colombia. In the development of their functions, the directors make decisions called orders, sentences or concepts. Once their decisions are firm, they are sent to the corporation's rapporteurship for analysis, standardization, titling and systematization. In this way they are available to the public for consultation. The titling work requires a work team, in charge of including in the system the general data of the orders, labeling them through keywords (descriptors and restrictors), extracting the thesis and the aspects that they consider relevant for the proper classification and subsequent decision consultation.

This project was divided into two stages. In the first stage, it seeks to reduce the mechanical work of standardizing the information embodied in the different documents (providence, orders, rulings) that the rapporteurs advance, through the use of Machine Learning techniques, since Artificial Intelligence (AI ), these data can be extracted automatically and in shorter periods of time than those used by the rapporteurs. This will allow the Rapporteurship team to carry out the legal tasks of their position, that is, the analysis and study of the corporation's decisions. In the Logical Framework approach, the central problem of the research lies in the difficulty of consulting the material of the Rapporteurship of the Council of State, due to the variety of terms and writing styles in which these documents are generated and stored. . Given the circumstances, the execution of a system to automate the certification process of the rapporteurships will facilitate public access to the documentation of the Council of State, through the classification and standardization of documents. In the second stage, the aim is to automate part of the process of identification and extraction of the theses and the legal problems of the sentences of the First Section of the Council of State. This work is proposed by means of an Intelligent System (IS) based on Natural Language Processing (NLP) for the analysis and interpretation of sentences, which facilitates, in an effective and efficient way, the understanding and automatic extraction of the Legal Problem and the Thesis.